On Monday, Morgan Stanley said in a new research note that the global hotel industry's distribution model could face a major overhaul as AI agents gradually enter travel search and hotel booking. AI may not only help hotels reduce their reliance on online travel agencies (OTAs), but also lift the share of bookings made directly through hotel websites through more personalized recommendations, thereby lowering the industry's overall marketing and distribution costs.

The bank believes international hotel groups with vast loyalty programs will be the main beneficiaries, and it has raised its earnings estimates and price targets for several hotel companies including Marriott International (MAR.US) and Hilton Worldwide (HLT.US).

Morgan Stanley estimates that global travel booking transactions exceed $1.6 trillion annually, of which hotel room revenue is about $640 billion. OTAs contribute about $220 billion of hotel booking transaction value and earn roughly $30 billion in related fee revenue from it.

At the same time, the global hotel industry spends about $22 billion to $55 billion on marketing each year, with about 40% to 50% of that going to search channels such as Google (GOOGL.US). Combining OTA commissions and hotel marketing expenses, Morgan Stanley estimates that AI agents could influence about $55 billion to $85 billion of related spending in the global hotel industry each year.

The bank notes that the traditional hotel booking process usually begins with a Google search, an OTA platform or a hotel's official website, while the rise of AI agents could change how consumers find and book hotels. In the future, consumers may describe their travel purpose, budget, accommodation preferences and loyalty benefits directly to AI, which would then screen hotels and recommend the most suitable booking channel.

This shift is expected to lower hotel customer acquisition costs, especially benefiting large hotel groups with mature loyalty programs and strong brand recognition. Morgan Stanley's latest research found that, compared with traditional search engines, large language models (LLMs) are clearly more inclined to show branded hotels and their official websites when recommending hotels.

According to tests by the bank's AlphaWise team across multiple US travel destinations, branded hotels accounted for about 41% of traditional Google search results while independent hotels accounted for about 59%. In recommendations generated by large language models, branded hotels rose to 87% while independent hotels fell to 13%.

The difference in booking links is even more striking. In US destination tests, hotel official website links made up only about 13% of Google search results, while OTA platform links accounted for about 78%. By contrast, hotel official website links reached as high as 97% in large language model recommendations. A similar trend appeared in Europe. Morgan Stanley's tests in London, Nice and Florence showed that hotel official website links accounted for 74% of large language model recommendations, above the roughly 8% seen in Google search results.

However, the bank stressed that these results reflect the current AI recommendation mechanism and do not mean all future hotel bookings will shift to direct channels. Morgan Stanley further found that hotel loyalty programs could become an important factor determining AI recommendation results. When consumers explicitly state in an AI search that they hold membership in a certain hotel group, AI recommendations clearly tilt toward the corresponding brand and its official website.

This means hotel groups' long-accumulated member data, points rewards, exclusive rates and membership tier benefits could translate into a stronger competitive advantage in the AI era. Morgan Stanley believes large chain hotel groups such as Marriott International (MAR.US), Hilton Worldwide (HLT.US) and Hyatt Hotels (H.US) are best positioned to use their vast loyalty systems, proprietary customer data and technology partnerships to drive more consumers to book directly.

By comparison, although smaller hotel groups could also benefit from lower distribution costs, their actual gains may fall short of larger competitors because of relatively limited loyalty program scale and brand premium.

A higher share of direct hotel bookings would also bring clear financial benefits. Morgan Stanley notes that OTA platforms typically charge hotels commissions of about 10% to 20%, while the related cost of completing a booking through a branded hotel's official website is about 5%. Therefore, even if only a small number of bookings shift from OTA platforms to hotel official websites, it could improve hotel operating margins.

The bank estimates that for every 1 percentage point increase in the share of direct hotel bookings, hotel owners' EBITDA could rise by about 0.1% to 0.5%, while hotel brand operators' EBITDA could rise by about 0.25% to 0.45%. In addition, for every 1 percentage point decline in OTA commission rates, hotel owners' and brand operators' EBITDA could gain about 0.15% to 0.65% and 0.25% to 0.65%, respectively.

Based on the above analysis, Morgan Stanley expects the distribution cost savings brought by AI agents to gradually appear starting in the second half of 2027 and further drive earnings improvement for hotel companies in 2028. The bank expects that by 2028, the average margin of hotel real estate investment trusts (REITs) under its coverage will be about 0.2 percentage points higher than previously forecast, while large hotel brand operators' franchise and related fee rates are expected to rise by about 0.1 percentage point on average.

On valuation, Morgan Stanley raised price targets for several US hotel companies. It lifted Marriott International (MAR.US) to $384 from $380, Hilton Worldwide (HLT.US) to $336 from $332, Hyatt Hotels (H.US) to $211 from $209, Wyndham Hotels & Resorts (WH.US) to $92 from $91, and Choice Hotels International (CHH.US) to $92 from $91.

For hotel REITs, Morgan Stanley's target price adjustments were even more notable. The bank raised Park Hotels & Resorts (PK.US) to $13 from $11, an increase of about 18%; DiamondRock Hospitality (DRH.US) to $12 from $11; Host Hotels & Resorts (HST.US) to $24 from $22; and Pebblebrook Hotel Trust (PEB.US) to $14 from $13. Morgan Stanley believes hotel REITs could directly benefit from lower distribution costs, so their valuation improvement potential may be greater than that of brand operators. The bank maintained its overweight ratings on Marriott International (MAR.US) and Hilton Worldwide (HLT.US).

However, the rise of AI agents does not mean online travel platforms such as Booking Holdings (BKNG.US) and Expedia (EXPE.US) will lose competitiveness. Morgan Stanley notes that OTA platforms have broad hotel inventory, mature loyalty systems, payment capabilities and order fulfillment infrastructure, advantages that remain difficult for AI agents to fully replace.

At the same time, AI may mainly change how consumers discover hotels rather than completely disrupt the final transaction stage. For consumers seeking distinctive accommodations or independent boutique hotels, the rich hotel resources of OTA platforms may instead play a greater role in AI recommendations. In addition, OTA platforms may also reduce their own marketing spending due to intensifying competition in search channels, offsetting part of the impact of lower commission revenue.

Therefore, Morgan Stanley's overall view on Booking Holdings (BKNG.US) and Expedia (EXPE.US) is relatively neutral, and it kept both companies' price targets unchanged. However, considering that large hotel chains may use AI to attract more consumers to book directly, the bank cut its 2027 and 2028 EBITDA forecasts for Expedia (EXPE.US) by 1% and 6%, respectively, while maintaining its $235 price target. Booking Holdings' (BKNG.US) earnings forecasts remained unchanged.

For Google (GOOGL.US), Morgan Stanley believes AI agents could pose some challenge to its traditional travel search advertising business. The bank estimates that a considerable portion of hotel-related marketing spending goes to search engines, and AI agents will become a new entry point for consumers seeking accommodation information. As travel search traffic shifts to large language models, Google (GOOGL.US) may face fiercer competition.

However, Morgan Stanley notes that travel-related business accounts for about 5% of Google's (GOOGL.US) revenue, and even if some revenue is affected, its overall AI business opportunity may still be far greater than the risk facing the travel search market.

It is worth noting that the bank also reminded investors that the long-term impact of AI agents on the hotel industry remains uncertain. If AI platforms strengthen commercialization in the future through advertising, sponsored recommendations or commission models, hotel distribution costs could rise again. At the same time, if a few large AI platforms eventually control the main travel search traffic, their bargaining power could even exceed that of traditional search engines and OTA platforms. In addition, AI agents may also directly take on booking and transaction functions in the future, further changing the distribution of interests among hotel groups, OTA platforms and technology companies.

Morgan Stanley believes that although the above risks deserve attention, current research results already show that AI agents could help large hotel groups increase direct booking shares and reduce the industry's persistently high distribution costs. The bank expects that as AI technology becomes further integrated with hotel loyalty programs and booking systems, hotel groups' brand influence and member loyalty will become even more important, and companies with large member bases and strong direct sales capabilities are expected to occupy a more favorable position in this round of industry change.